Corticosteroids have been a key element of acute lymphocytic leukemia (ALL) induction therapy for decades, yet it has been challenging to reach definitive conclusions about the optimal steroid formulation. Several recent clinical trials in B- and T-lineage ALL have compared dexamethasone versus prednisone during the induction phase of treatment.1 The rationale for using dexamethasone includes greater potency and central nervous system (CNS) penetration, which is particularly appealing in T-ALL given the higher rates of CNS disease. This has been counterbalanced, however, by higher rates of infectious toxicity and treatment-related mortality with dexamethasone-based induction regimens. Past studies have determined that selecting the optimal steroid is complex, and unique features of the patient population and the other components of the treatment regimen influence the effect of induction corticosteroids. Thus, an overarching consensus about the optimal induction steroid has not been reached.
The AIEOP-BFM ALL 2000 trial conducted by the Associazione Italiana di Ematologia e Oncologia Pediatrica (AIEOP) in collaboration with Berlin-Frankfurt-Münster (BFM) ALL study groups from 2000 to 2006 revisited this question with a randomization between dexamethasone vs. prednisone during the induction phase to determine whether treatment with dexamethasone provides a better event-free and overall survival in newly diagnosed childhood B-ALL and T-ALL. Pediatric patients aged one to 17 years participated in this trial at 127 participating sites in Europe. After receiving a seven-day prednisone prephase, children were randomized to receive either prednisone (60 mg/m2/day) or dexamethasone (10 mg/m2/day) for an additional 21 days during a four-drug induction with a subsequent taper. A total of 3,720 patients were randomized: 1,853 were assigned to the dexamethasone arm, and 1,867 patients to the prednisone arm.
Post-induction, children continued treatment with multiagent chemotherapy for a total of two years or with hematopoietic allogeneic stem cell transplantation in first remission for those with very-high-risk features. Prednisone good and poor responses were defined as <1.0 × 109/L or ≥1.0 ×109/L peripheral blood blasts, respectively, after the seven-day prednisone prephase and one intrathecal dose of methotrexate. The primary outcome in this study was event-free survival, with events defined as nonresponse, relapse, secondary neoplasm, or death from any cause. The steroid randomization was halted in October 2004 for patients aged 10 years or older, due to safety concerns.
Strikingly, the overall relapse incidence was reduced by one-third in the dexamethasone arm (10.8% ± 0.7% vs. 15.6% ± 0.8%; p<0.0001), with a proportionally greater reduction in extramedullary versus early bone marrow relapses. However, treatment-related death rates during induction were significantly higher in the dexamethasone arm (2.5% vs. 0.9%; p=0.00013). Among patients ≥ 10 years of age, the reduction in relapse risk was offset by this higher incidence of toxic deaths during induction such that no differences in event-free survival were observed in this older age group. Five-year event-free survival was significantly better for patients randomized to the dexamethasone arm compared with patients assigned to receive prednisone (83.9% ± 0.9% vs. 80.8% ± 0.9%; p=0.024, respectively). However, no difference between the randomization groups could be demonstrated for five-year overall survival (dexamethasone, 90.3% ± 0.7%; prednisone, 90.5% ± 0.7%).
Among prednisone poor responders, there were no differences in event-free or overall survival or relapse incidence in the randomized groups for either B- or T-lineage ALL. Among the patients with a prednisone good response, a significantly lower incidence of relapse and better event-free survival was observed in the dexamethasone arm compared to the prednisone arm for patients with both B- and T-lineage ALL. In the T-ALL patients with a prednisone good response, relapse rates were notably reduced from 17 percent to 7 percent with dexamethasone as the induction steroid and the better event-free survival of the dexamethasone group also translated into significantly better survival of 91.4% ± 2.4% compared to 82.6% ± 3.2%; p=0.036 with prednisone. This was in contrast to the patients with prednisone good-response and B-ALL, who had an even inferior, though not statistically significant, survival rate in the dexamethasone group, largely attributable to inferior survival rates following relapse (see Table).
Table. A Comparison of Dexamethasone vs. Prednisone During Induction
| . | Dexamethasone . | Prednisone . | P Value . |
|---|---|---|---|
| 5-year cumulative incidence of relapse | 10.8% ± 0.7% | 15.6% ± 0.8% | < 0.001 |
| Treatment-related deaths in induction | 2.5% | 0.9% | 0.0013 |
| 5-year EFS | 83.9% ± 0.9% | 80.8% ± 0.9% | 0.24 |
| 5-year OS | 90.3% ± 0.7% | 90.5% ± 0.7% | |
| 5-year OS in PGR T-ALL | 91.4% ± 2.4% | 82.6% ± 3.2% | 0.036 |
| 5-year survival after relapse in PGR T-ALL | 36.4% ± 9.5% | 23.1% ± 8.3% | 0.62 |
| 5-year OS in PGR B-ALL | 91.9% ± 0.7% | 93.4% ± 0.6% | 0.12 |
| 5-year survival after relapse in PGR B-ALL | 51.9% ± 4.1% | 65.7% ± 3.1% | 0.0053 |
| . | Dexamethasone . | Prednisone . | P Value . |
|---|---|---|---|
| 5-year cumulative incidence of relapse | 10.8% ± 0.7% | 15.6% ± 0.8% | < 0.001 |
| Treatment-related deaths in induction | 2.5% | 0.9% | 0.0013 |
| 5-year EFS | 83.9% ± 0.9% | 80.8% ± 0.9% | 0.24 |
| 5-year OS | 90.3% ± 0.7% | 90.5% ± 0.7% | |
| 5-year OS in PGR T-ALL | 91.4% ± 2.4% | 82.6% ± 3.2% | 0.036 |
| 5-year survival after relapse in PGR T-ALL | 36.4% ± 9.5% | 23.1% ± 8.3% | 0.62 |
| 5-year OS in PGR B-ALL | 91.9% ± 0.7% | 93.4% ± 0.6% | 0.12 |
| 5-year survival after relapse in PGR B-ALL | 51.9% ± 4.1% | 65.7% ± 3.1% | 0.0053 |
EFS, event-free survival; OS, overall survival; PGR, prednisone good response; T-ALL, T-cell acute lymphocytic leukemia; B-ALL, B-cell acute lymphocytic leukemia.
The results of this trial have several important implications for the treatment of B- and T-lineage childhood ALL. The benefits of dexamethasone in reducing relapse rates by one third were notable given that the induction steroid was the only change in an intensive multi-agent treatment regimen; this intervention has had the largest impact on relapse risk reduction in decades. This benefit, however, was offset by toxicity and treatment-related mortality, especially among patients 10 years or older, which is consistent with the experience of other groups.2 Any benefits of dexamethasone also seemed to be absent in prednisone poor responders, suggesting that a more potent steroid formulation may not be enough to overcome the impact of a steroid-resistant phenotype.
In Brief
The lack of survival benefit in B-ALL, despite an improvement in event-free survival is another key finding, which could be explained by the observation that dexamethasone had the greatest impact in preventing B-ALL relapses that were more readily salvageable, such as extramedullary or later recurrences. These observations highlight the importance of evaluating the extent to which event-free survival benefits translate into improvements in overall survival as well. One of the most notable finings in this report is the benefit for dexamethasone in T-ALL patients with prednisone good responses. Within this subgroup, relapse rates were markedly reduced and event-free and overall survival improved significantly with dexamethasone as the induction steroid. Notably, overall survival rates exceeded 90 percent in this population that has been challenging to treat and salvage historically. Going forward, the results from this trial continue to suggest that there is not a straightforward answer as to whether one steroid is universally superior to another during childhood ALL induction therapy; rather, there are selective benefits to different formulations among patient subsets and a personalized approach appears warranted, with emphasis on preventing relapses that cannot be salvaged.
References
Competing Interests
Dr. Raetz indicated no relevant conflicts of interest.
